Taxonomic Structure and Monitoring of the Dominant Population of Lactic Acid Bacteria during Wheat Flour Sourdough Type I Propagation Using Lactobacillus sanfranciscensis Starters

Abstract

The structure and stability of the dominant lactic acid bacterium population were assessed during wheat flour sourdough type I propagation by using singly nine strains of Lactobacillus sanfranciscensis. Under back-slopping propagation with wheat flour type 0 F114, cell numbers of presumptive lactic acid bacteria varied slightly between and within starters. As determined by randomly amplified polymorphic DNA-PCR and restriction endonuclease analysis-pulsed-field gel electrophoresis analyses, only three (LS8, LS14, and LS44) starters dominated throughout 10 days of propagation. The others progressively decreased to less than 3 log CFU g⁻¹. Partial sequence analysis of the 16S rRNA and recA genes and PCR-denaturating gradient gel electrophoresis analysis using the rpoB gene allowed identification of Weissella confusa, Lactobacillus sanfranciscensis, Lactobacillus plantarum, Lactobacillus rossiae, Lactobacillus brevis, Lactococcus lactis subsp. lactis, Pediococcus pentosaceus, and Lactobacillus spp. as the dominant species of the raw wheat flour. At the end of propagation, one autochthonous strain of L. sanfranciscensis was found in all the sourdoughs. Except for L. brevis, strains of the above species were variously found in the mature sourdoughs. Persistent starters were found in association with other biotypes of L. sanfranciscensis and with W. confusa or L. plantarum. Sourdoughs were characterized for acidification, quotient of fermentation, free amino acids, and community-level catabolic profiles by USING Biolog 96-well Eco microplates. In particular, catabolic profiles of sourdoughs containing persistent starters behaved similarly and were clearly differentiated from the others. The three persistent starters were further used for the production of sourdoughs and propagated by using another wheat flour whose lactic acid bacterium population in part differed from the previous one. Also, in this case all three starter strains persisted during propagation.

Based on common principles used in artisanal and industrial processes, three types (I, II, and III) of sourdough have been arbitrarily defined (4). Type I sourdough has the largest application and corresponds to traditional processes. It is characterized by continuous (daily) propagation to keep the microorganisms in an active state, as indicated by high metabolic activity. Propagation is achieved either by back-slopping, using the mother sponge taken from the preceding fermentation process, or by using commercial starters once a week, followed by daily back-slopping (15). Obviously, the second option is largely preferred for industrial processes since it is less laborious and more standardized, especially if the use of starters could be extended monthly.

When sourdough is continuously propagated, a stable association of only few species of lactic acid bacteria and yeasts might be achieved to ensure a controlled process. Under these conditions, the microbial ecology of sourdough is influenced by such factors as the chemical, enzyme, and microbial composition of cereals, environmental conditions (e.g., temperature, pH, and redox potential), and technology (e.g., process parameters and use of starters and baker's yeast) (15, 21). As a consequence of the heterogeneity of the above factors, a large microbial diversity is generally found between mature sourdoughs. Several new species have recently been isolated from traditional sourdoughs, and almost 50 species belonging to the genera Leuconostoc, Pediococcus, Enterococcus, Weissella, and especially Lactobacillus have been enumerated (15). Although lactic acid bacteria isolated from sourdough are not necessarily unique for sourdough ecosystems, correlations between species and types (I, II, or III) of sourdough have been found. Lactobacillus sanfranciscensis seems to be the key lactic acid bacterium in sourdough type I (5, 18, 21, 23, 25, 29, 32). Contrary to the case with most of the other sourdough lactobacilli, which are also inhabitants of other ecosystems (e.g., intestines of humans and animals), no other habitats are currently known for L. sanfranciscensis (21). Almost all the properties attributed to sourdough have been attributed to the activity of L. sanfranciscensis (19). It positively contributes to heterolactic acidification, flavor, proteolysis, texture, and nutritional properties of sourdough baked goods.

Once selected based on technology and functional features, starters have to be used in the complex sourdough matrix, and especially, they have to dominate during back-slopping processes. Compared to other food or beverage fermentations (e.g., cheese and wine making), bakery industries enjoy great advantages by using stable sourdough starters (e.g., monthly) due to the low market prices of most baked goods. Several studies have dealt with the stability of the bacterial communities in sourdough ecosystems. A large number of traditional sourdoughs were sampled twice at 1-year intervals, in 11 bakeries located throughout Belgium (32). Microbial stability in two bakery sourdoughs made from conventionally and organically grown rye flour was studied (29), and the biodiversity of L. sanfranciscensis was investigated in sourdough samples from the United States and Germany (23). The evolution of cereal grain-related subdominant sourdough lactic acid bacteria was monitored during laboratory-scale fermentations in the presence of selected L. sanfranciscensis strains (7). Only one study monitored four sourdoughs produced under practical conditions by using a starter mixture of three commercially available sourdough starters and baker's yeast (25). Although L. sanfranciscensis emerged as one of the predominant species in almost all the sourdoughs, results regarding the source of the dominant strains (autochthonous or starter) and the factors (cereals and/or technology) influencing the stable association were controversial.

This study describes the isolation, identification, monitoring, and technology performance of the dominant lactic acid bacterium population during wheat flour sourdough type I propagation by using L. sanfranciscensis starters singly.

MATERIALS AND METHODS

Microorganisms and growth conditions.

Lactobacillus sanfranciscensis LS3, LS6, LS8, LS12, LS14, LS40, LS41, LS44, and LS48, isolated from Italian sourdoughs and belonging to the Culture Collection of the Department of Plant Protection and Applied Microbiology of Bari University, Bari, Italy, were used in this study. All strains had been identified by partial 16S rRNA gene sequencing and subjected to technological and functional characterization (12). Cell cultures were routinely propagated at 30°C for 24 h in SDB broth (24). Growth was monitored directly by plating on SDB agar during 24 h of incubation at 30°C or indirectly by measuring the optical density at 620 nm (OD₆₂₀). When used for sourdough fermentation, cells of lactobacilli were incubated until the exponential phase of growth was reached (ca. 12 h) (cells diluted 1:10 gave an OD₆₂₀ of ca. 0.25), harvested by centrifugation (9,000 × g for 10 min at 4°C), washed once with 50 mM phosphate buffer, pH 7.0, and resuspended in sterile distilled water at a cell density of about 9 log CFU ml⁻¹ (cells diluted 1:10 gave an OD₆₂₀ of ca. 0.30).

Laboratory sourdough production and propagation.

Two type 0 Triticum aestivum flours (F114 and F164) were used. Moisture was determined according to the standard AACC method (2). The nitrogen content was measured by using the semimicro-Kjeldahl method. Protein was calculated by multiplying the nitrogen concentration (N) by 5.7 (3, 12). The total concentration of starch was determined using the Megazyme starch assay kit (Megazyme International, Wicklow, Ireland), official method 76-13 (3). Carbohydrates were determined by high-performance liquid chromatography (39). Starch damage in flour was determined according to the Megazyme AACC method (3), using the Megazyme starch damage kit (Megazyme International Ireland Ltd., Bray, Ireland). The gross composition of F114 and F164 flours was as follows: moisture, 15.65% ± 0.21% and 15.02% ± 0.67%; protein (N × 5.70), 12.32% ± 0.02% and 11.55% ± 0.01% of dry matter; starch, 67.35% ± 0.47% and 68.84% ± 0.39% of dry matter; total soluble carbohydrates, 3.87% ± 0.38% and 3.95% ± 0.03% of dry matter; and damaged starch, 6.69% ± 0.33% and 5.10% ± 0.50% of dry matter, respectively. Two types of wheat flour were used to confirm the abilities of some starters to dominate the indigenous microbiota.

Sourdough production and propagation were established based on the most diffuse and traditional protocol used for bread making in the central and southern areas of Italy. The production of sourdoughs was carried out according to the following protocol. Wheat flour F114 (187.5 g), tap water (82.5 ml), and cell suspension (30 ml), containing about 9 log CFU ml⁻¹ of the above individual L. sanfranciscensis strains (final cell number in the dough of ca. 7.7 log CFU g⁻¹ of dough), were used to prepare 300 g of dough [dough yield (dough weight × 100)/(flour weight) = 160] with a continuous high-speed mixer (60 × g; dough mixing time, 5 min) (Chopin & Co., Boulogne, Seine, France). A control sourdough, without starter, was also produced under the same conditions. The 10 sourdoughs (9 with individual L. sanfranciscensis strains and the control) were incubated in sterile plastic beakers at 30°C for 8 h. After fermentation, sourdoughs were stored at 10°C for about 16 h and further used for propagation.

Sourdough propagation was carried out according to the back-slopping protocol. Each of the above sourdoughs was individually used as the starter (inoculum rate, 10% [wt/wt]) to ferment a mixture of flour F114 (168.75 g) and tap water (101.25 ml) having a dough yield of 160. Sourdoughs were fermented at 30°C for 6 h and stored at 10°C for about 16 h between back-slopping. Sourdoughs were propagated daily for 10 days.

Lactic acid bacterium and yeast enumeration and isolation.

Ten grams of sourdough were homogenized with 90 ml of sterile sodium chloride (0.9% [wt/vol]) solution using a Classic Blender homogenizer (PBI International, Milan, Italy). Serial dilutions were plated on SDB agar for determination of presumptive lactic acid bacteria. SDB agar containing 1% (wt/vol) of α-cyclodextrin, α-methyl-d-glucoside and l-methionine as the only carbon sources (12) was used to enumerate the starter strains of L. sanfranciscensis. After incubation at 30°C for 48 h, at least 10 colonies were randomly selected from plates containing the two highest sample dilutions. Isolates that were gram-positive, catalase-negative, nonmotile rod and cocci were cultivated in SDB broth at 30°C for 24 h and restreaked into SDB agar. All isolates considered for further analyses could acidify the culture medium. Stock cultures were stored at −20°C in 10% (vol/vol) glycerol.

The number of yeasts in sourdoughs was estimated by using malt agar (Difco, Detroit, MI) at 30°C for 72 h. Randomly selected colonies of yeast were subcultured on yeast malt broth (Difco) and restreaked into malt agar (29). Identification of yeasts was carried out by using Biolog 96-well yeast (YT) microplates (Biolog, Inc., Hayward, CA) (27).

DNA extraction and molecular identification by 16S rRNA and recA gene sequencing.

Total DNA from presumptive lactic acid bacterium isolates was extracted from 2-ml samples of overnight cultures grown at 30°C in SDB broth. Total DNAs were obtained as described by de Los Reyes-Gavilán et al. (13). The concentration and purity of DNA were assessed by using a biophotometer (Eppendorf, Hamburg, Germany). A primer pair (Invitrogen Life Technologies, Milan, Italy), LpigF/LpigR (5′-TACGGGAGGCAGCAGTAG-3′ and 5′-CATGGTGTGACGGGCGGT-3′ [11]) (corresponding to positions 369 to 386 and 1424 to 1441, respectively, of the 16S rRNA gene sequence of Lactobacillus mucosae, accession number AF126738), was used to amplify the 16S rRNA gene fragment of presumptive lactic acid bacteria. Fifty microliters of each PCR mixture contained 200 μM of each 2′-deoxynucleoside 5′-triphosphate, 1 μM of both forward and reverse primer, 2 mM MgCl₂, 2 U of Taq DNA polymerase (Invitrogen) in the supplied buffer, and approximately 50 ng of DNA. PCR amplification was performed using the GeneAmp PCR system 9700 thermal cycler (Applied Biosystems). Amplification consisted of an initial denaturation step at 94°C for 4 min, followed by 30 cycles at 94°C for 45 s at 49°C for 45 s and a final extension at 72°C for 1 min. An additional final extension at 72°C for 7 min was used. PCR products were separated by electrophoresis on a 1.5% (wt/vol) agarose gel (Gibco BRL, France) stained with ethidium bromide (0.5 μg/ml). The amplicons were eluted from the gel and purified by the GFX PCR DNA and gel band purification kit (GE Healthcare Life Sciences, Milan, Italy). DNA sequencing reactions were performed by MWG Biotech AG (Ebersberg, Germany) using both forward and reverse primers. Taxonomic identification of strains was performed by comparing the sequences of each isolate with those reported in the Basic BLAST database (1) (http://www.ncbi.nlm.nih.gov).

Primers designed on the recA gene were also used to discriminate between Lactobacillus plantarum, Lactobacillus pentosus, and Lactobacillus paraplantarum species (36). Part of the recA gene was amplified using the degenerate primer pair (MWG Biotech AG, Ebersberg, Germany) recALb1F/recALb1R (5′-CRRTBATGCGBATGGGYG-3′ and 5′-CGRCCYTGWCCAATSCGRTC-3′), derived from the homologous regions of the recA gene sequences from Lactobacillus plantarum subsp. plantarum (accession no. AJ621668). PCRs were performed as described for the 16S rRNA gene, and related amplicons were separated, purified, and sequenced as described above.

rpoB gene analysis by PCR-denaturating gradient gel electrophoresis (DGGE) was performed to discriminate between Weissella cibaria and Weissella confusa species using the type strains (W. cibaria LMG 17699^T and W. confusa JCM 1093^T). The analysis was performed as described by De Angelis et al. (12).

Genotypic characterization by RAPD-PCR analysis.

Genomic DNA from each isolate was extracted as described above. Three oligonucleotides, P4 (5′-CCGCAGCGTT-3′), P7 (5′-AGCAGCGTGG-3′) (6), and M13 (5′-GAGGGTGGCGGTTCT-3′) (35), with arbitrarily chosen sequences, were used for isolate biotyping. The reaction mixture and PCR conditions reported by Corsetti et al. (6) were used for primers P4 and P7, whereas those reported by Zapparoli et al. (41) were used for primer M13. The PCR products (15 μl) were separated by electrophoresis at 100 V for 200 min on a 1.5% (wt/vol) agarose gel, and the DNA was detected by UV transillumination after staining with ethidium bromide (0.5 μg ml⁻¹). The molecular sizes of the amplified DNA fragments were estimated by comparison with the 1-kb DNA molecular size markers (Invitrogen Life Technologies). Randomly amplified polymorphic DNA-PCR (RAPD-PCR) profiles were acquired by using the Gel Doc EQ system (Bio-Rad, Hercules, CA) and compared using the Fingerprinting II Informatix software program (Bio-Rad). The similarity of the electrophoretic profiles was evaluated by determining the Dice coefficients of similarity and by using the unweighted-pair group method using average linkages. The presence or absence of fragments was recorded as 1 or 0, respectively. Only reproducible well-marked amplified fragments were scored, with faint bands being ignored. The reproducibility of the RAPD fingerprints was assessed by comparing the PCR products obtained with primers P4, P7, and M13 and DNA prepared from three separate cultures of the same strain.

Genotypic characterization by restriction endonuclease analysis-pulsed-field gel electrophoresis (REA-PFGE).

Intact genomic DNA was isolated as previously described by Moschetti et al. (26). DNA inserts were digested in 200 μl of appropriate buffer supplemented with 40 U of SmaI or ApaI (Promega Co., Madison, WI). Electrophoresis of the restriction digests was performed by using the Chef system (Bio-Rad Laboratories, Hercules, CA) with 1% (wt/vol) agarose gels and 0.5× Tris-borate-EDTA as the running buffer at 10°C. Restriction fragments were resolved in a single run at a constant voltage of 6 V cm⁻² and an orientation angle of 120° between electric fields by using a single-phase procedure for 20 h with a pulse ramping between 1 and 10 s.

Acidification during sourdough propagation.

Acidification during sourdough propagation was monitored online using a pH meter (model 507; Crison, Milan, Italy) with a food penetration probe. Concentrations of organic acids were determined by high-performance liquid chromatography using ÄKTApurifier systems (GE Healthcare Life Sciences). The analyses were carried out isocratically at 0.8 ml min⁻¹ at 65°C with a 300- by 7.8-mm-inside-diameter cation exchange column (Aminex HPX-87H) (Bio-Rad Laboratories). The mobile phase was 0.013 N H₂SO₄, prepared by diluting reagent-grade sulfuric acid with distilled water, filtering through a 0.45-μm-membrane filter (Sartorius, AG, Göttingen, Germany), and degassing under vacuum. A mixture of organic acids at known concentrations (Sigma Chemical Co.) was added and used as a standard (42). The quotient of fermentation (QF) (lactic acid/acetic acid) was determined as the molar ratio between d,l-lactic and acetic acids (12).

Determination of free amino acids.

The concentrations of free amino acids (FAA) in water extracts of sourdoughs were determined by using an amino acid analyzer (Biochrom Ltd., Cambridge Science Park, England) using a Na cation exchange column (20- by 0.46-cm inside diameter). A mixture of amino acids at known concentrations (Sigma Chemical Co.) was added with cysteic acid, methionine sulfoxide, methionine sulfone, tryptophan, and ornithine and used as a standard. Proteins and peptides in the samples were precipitated by addition of 5% (vol/vol) cold solid sulfosalicylic acid, holding at 4°C for 1 h and centrifuging at 15,000 × g for 15 min. The supernatant was filtered through a 0.22-μm-pore-size filter and diluted, when necessary, with a sodium citrate (0.2 M, pH 2.2) loading buffer. Amino acids were postcolumn derivatized with ninhydrin reagent and detected by measuring the absorbance at 440 (proline and hydroxyproline) or 570 nm (all the other amino acids) (12).

Community-level catabolic profiles (CLCPs).

Carbon source utilization patterns of sourdough microbial communities during propagation were assessed by using Biolog 96-well Eco microplates (Biolog, Inc., Hayward, CA) (8). They contained 31 different carbon sources (carbohydrates, carboxylic acids, polymers, amino acids, amines, and miscellaneous substrates) in triplicate. Ten grams of sourdough was homogenized with 90 ml of sterile sodium chloride (0.9% [wt/vol]) solution (Classic Blender) and centrifuged at 10,000 × g for 15 min at 4°C. The pellet was washed with 50 mM Tris-HCl (pH 8.8) and then with sterile sodium chloride solution and centrifuged as described above. The cellular suspension was diluted (1:100) in sterile sodium chloride solution and subsequently dispensed (150 μl) into each of 96 wells of the Biolog Eco microplates. The microplates were incubated at 30°C in the dark, and the color development was measured at 590 nm every 24 h thereafter up to 120 h with a microplate reader (Biolog Microstation). Three indices were determined (33, 34, 40). Shannon's diversity (H′), indicating the substrate utilization pattern was calculated as follows: H′ = −Σ p_i ln (p_i), where p_i is the ratio of the activity of a particular substrate to the sums of activities of all substrates at 120 h. Substrate richness (S), measuring the number of different substrates used, was calculated as the number of wells with a corrected absorbance greater than 0.25. Substrate evenness (E) was defined as the equitability of activities across all utilized substrates: E = H′/log S.

Statistical analysis.

Data (three replicates) from microbial growth, acidification, organic acids, and FAA were subjected to one-way analysis of variance (30), and pair comparison of treatment means was achieved by using Tukey's procedure at P values of <0.05 (37), using the statistical software Statistica for Windows (Statistica 6.0 per Windows 1998). In addition, some data were further subjected to nonparametric statistical analysis by determining the median value of aggregate strain data.

RESULTS

Lactic acid bacterium enumeration and isolation.

Preliminarily, nine strains of Lactobacillus sanfranciscensis were individually used to produce as many sourdoughs made of wheat flour type 0 F114. After 8 h of fermentation at 30°C, all sourdoughs had pH values in the range of 3.95 to 3.80 and contained 8.9 to 9.5 log CFU g⁻¹ of presumptive L. sanfranciscensis. After 8 h of fermentation, the control sourdough contained a number of presumptive lactic acid bacteria about 10,000 times lower than that of the started sourdoughs (data not shown). Each sourdough was then propagated for 10 days by daily back-slopping. Cell densities after each sourdough fermentation are shown in Fig. 1. The median value of presumptive lactic acid bacteria in the control was 9.15 log CFU g⁻¹, and the values corresponding to 5th and 95th percentiles of the data were 8.18 and 9.58 log CFU g⁻¹, respectively. The lowest value (outlier, ca. 7.4 log CFU g⁻¹) was found after the first day of propagation. Cell numbers rapidly increased to 9.3 log CFU g⁻¹ during a further 3 days and remained almost constant. Cell densities of presumptive lactic acid bacteria in the sourdough started with L. sanfranciscensis LS3 (S-LS3) ranged from 8.96 to 9.9 log CFU g⁻¹, with a median value of 9.46 log CFU g⁻¹. Similar median values (9.02 to 9.36 log CFU g⁻¹) were found for the sourdoughs S-LS8, S-LS12, S-LS14, S-LS40, S-LS41, S-LS44, and S-LS48. The lowest median value was found for the sourdough started with L. sanfranciscensis LS6 (S-LS6). The narrowest ranges of the 5th and 95th percentiles, outlier and extreme of the aggregate data, were found for S-LS8 (8.97 to 9.49 log CFU g⁻¹), S-LS12 (9.00 to 9.23 log CFU g⁻¹), S-LS14 (8.68 to 9.36 log CFU g⁻¹), LS40 (9.0 to 9.58 log CFU g⁻¹), and S-LS44 (9.04 to 9.76 log CFU g⁻¹).

FIG. 1.

Cell densities (CFU g⁻¹) after sourdough propagation at 30°C for 6 h in 10 subsequent days with use of wheat flour type 0 F114. Data are the means from three independent experiments (n = 3). The center line of each box (□) represents the median; the top and bottom of the box represent the 75th and 25th percentiles of the data, respectively. The tops and bottoms of the error bars represent the 5th and 95th percentiles of the data, respectively. The circles (○) in each box plot extend to the outliers of the data, and very extreme points are represented as individual data points (*). Control, unstarted sourdough. Sourdough (S) started with strain Lactobacillus sanfranciscensis LS3, LS6, LS8, LS12, LS14, LS40, LS41, LS44, or LS48 is indicated.

Apart from the type of sourdough, yeasts reached cell densities of about 7 log CFU g⁻¹ since the fourth day of propagation and remained almost constant. Phenotypic identification of isolates by Biolog 96-well yeast (YT) microplates indicated Saccharomyces cerevisiae as the dominant species in all sourdoughs (data not shown).

Dynamics of the lactic acid bacteria population during sourdough propagation.

Isolates of presumptive lactic acid bacteria from each daily sourdough propagation were subjected to RAPD-PCR analysis by using the single primer P4, P7, or M13. The patterns showed about 92 to 95% similarity, indicating the reproducibility of the technique under the conditions used (data not shown). The use of primer P7 clearly differentiated the RAPD patterns of strains used as starters from those of all the other isolates either from sourdoughs or from the control (data not shown). The cell number of each starter was calculated as the number of colonies with the specific RAPD pattern. As expected, strains of L. sanfranciscensis dominated at the beginning of propagation (Fig. 2). The strains LS8, LS14, and LS44 maintained elevated cell numbers (ca. 9 log CFU g⁻¹) throughout 10 days of propagation. On the contrary, the other starters progressively decreased after the first day of propagation. After eight days, L. sanfranciscensis LS3, LS6, LS12, LS40, LS41, and LS48 almost disappeared (≤3 log CFU g⁻¹).

FIG. 2.

Kinetics of growth of Lactobacillus sanfranciscensis LS3 (▪), LS6 (⧫), LS8 (•), LS12 (*), LS14 (□), LS40 (▴), LS41 (…), LS44 (○), and LS48 (⋄) after sourdough propagation at 30°C for 6 h during 10 subsequent days with use of wheat flour type 0 F114. The data are the means from three independent experiments ± standard deviations (n = 3).

Lactic acid bacteria dominating the microbiota of raw wheat flour type 0 F114 were identified by sequence analysis of at least 700 bp of the 5′ region of the 16S rRNA gene. The partial sequence of the recA gene was also used to discriminate between Lactobacillus plantarum, Lactobacillus pentosus, and Lactobacillus paraplantarum. Based on 16S rRNA and recA gene analyses, the following species were found: Weissella cibaria/Weissella confusa (18 isolates), L. sanfranciscensis (2 isolates), L. plantarum (1 isolate), Lactobacillus rossiae (1 isolate), Lactobacillus brevis (1 isolate), Lactococcus lactis subsp. lactis (1 isolate), Pediococcus pentosaceus (2 isolates), and Lactobacillus spp. (3 isolates). The DGGE analysis of the PCR amplicons targeting a region of the gene coding for the RNA polymerase beta subunit (rpoB) was used to differentiate isolates of Weissella cibaria/W. confusa. Isolates were allotted to W. confusa by comparing their DGGE profiles with the corresponding profiles of the type strains (data not shown). The same analyses were carried out to identify isolates dominating the microbiota of the control and sourdoughs during propagation. The species identified after 10 days are shown in Table 1. W. confusa was found in the control and the sourdoughs S-LS6, S-LS14, S-LS41, and S-LS48; L. sanfranciscensis in all sourdoughs; Lactobacillus paralimentarius in S-LS6; L. rossiae in S-LS12, S-LS40, and S-LS41; L. plantarum in S-LS44; and P. pentosaceus and Lactococcus lactis subsp. lactis in S-LS48. L. brevis was the only species from the microbiota of raw wheat flour type 0 F114 that did not become dominant during propagation.

TABLE 1.

Identification by 16S rRNA gene sequencing of sourdough lactic acid bacteria^a

Sourdoughb	Isolatesc	Closest relative	Accession no. of closest relative (identity ≥ 98%)
Control	I(1)-I(4)	W. cibaria/ W. confusad	AJ295989/AB023241
	I(5)-I(10)	L. sanfranciscensis	EU350220
S-LS3	I(11)-I(20)	L. sanfranciscensis	EU350220
S-LS6	I(21)-I(26), I(28)	W. cibaria/ W. confusad	AJ295989/AB023241
	I(27), I(30)	L. sanfranciscensis	EU350220
	I(29)	L. paralimentarius	AJ417500
S-LS8	I(31), I(32)	L. sanfranciscensis	EU350220
	I(33)-I(40)	L. sanfranciscensis	EU350220
S-LS12	I(41), I(44)	L. rossiae	AJ564009
	I(42), I(43), I(45)-I(50)	L. sanfranciscensis	EU350220
S-LS14	I(51), I(53), I(57)-I(60)	L. sanfranciscensis	EU350220
	I(52), I(54), I(55)	L. sanfranciscensis	EU350220
	I(56)	W. cibaria/ W. confusad	AJ295989/AB023241
S-LS40	I(61), I(63), I(65)-I(70)	L. sanfranciscensis	EU350220
	I(62), I(64)	L. rossiae	AJ564009
S-LS41	I(71)	L. rossiae	AJ564009
	I(72)-I(74), I(76)-I(80)	W. cibaria/ W. confusad	AJ295989/AB023241
	I(75)	L. sanfranciscensis	EU350220
S-LS44	I(81)-I(83)	L. sanfranciscensis	EU350220
	I(84), I(87)-I(90)	L. sanfranciscensis	EU350220
	I(85)	L. plantarum	EF439680
	I(86)	L. plantarum	EF439682
S-LS48	I(91), I(92), I(95)	P. pentosaceus	AJ305321
	I(93), I(100)	L. sanfranciscensis	EU350220
	I(94)	L. lactis subsp. lactis	EU483103
	I(96)	P. pentosaceus	AJ305321
	I(97)-I(99)	W. cibaria/ W. confusad	AJ295989/AB023241

^aBacteria were isolated after 10 days of daily sourdough propagation at 30°C for 6 h with wheat flour type 0 F114.

^bLS3, LS6, LS8, LS12, LS14, LS40, LS41, LS44, and LS48 refer to strains of Lactobacillus sanfranciscensis used as the starters.

^cIsolates with different RAPD-PCR profiles are grouped in separate rows.

^dIsolates were allotted to W. confusa by comparing their DGGE profiles obtained after rpoB gene PCR with the corresponding profiles of the type strains.

Figure 3A shows the RAPD-PCR analysis (primer P7) of the different isolates from sourdoughs after 10 days of propagation. Sourdough S-LS3 contained isolates belonging to L. sanfranciscensis having a unique RAPD pattern. The same strain of L. sanfranciscensis was found in all sourdoughs (Table 1 and Fig. 3). The other sourdoughs harbored two or more species (e.g., two in the control and four in S-LS48) with a different number of dominant strains (e.g., five in S-LS48). In several cases, strains belonging to the same species had the same RAPD profiles in different sourdoughs. The same RAPD type of W. confusa was found in the control, S-LS6, S-LS14, S-LS41, and S-LS48, while the same RAPD type of L. rossiae was found in S-LS12, S-LS40, and S-LS41. On the contrary, sourdoughs S-LS44 and S-LS48 harbored two different biotypes of L. plantarum and P. pentosaceus, respectively. The isolates of L. sanfranciscensis found in S-LS8, S-LS14, and S-LS44 had the same RAPD pattern as the starters LS8, LS14, and LS44, respectively. Cluster analysis by using RAPD-PCR profiles from the primers P4, P7, and M13 of different isolates after 10 days of propagation confirmed the above results (Fig. 3B).

FIG. 3.

(A) Representative RAPD-PCR patterns of lactic acid bacteria isolated after 10 days of daily sourdough propagation at 30°C for 6 h with use of wheat flour type 0 F114. Primer P7 was used for RAPD-PCR analysis. st, DNA molecular size standards (12,000 to 100 bp); Control, unstarted sourdough. Sourdough started with Lactobacillus sanfranciscensis LS3 (S-LS3), LS6 (S-LS6), LS8 (S-LS8), LS12 (S-LS12), LS14 (S-LS14), LS40 (S-LS40), LS41 (S-LS41), LS44 (S-LS44), or LS48 (S-LS48) is indicated. Lactic acid bacterium isolates (I) are coded based on partial 16S rRNA and recA gene sequence comparisons and correspond to those of Table 1. Lb., Lactobacillus; Lc., Lactococcus. (B) Dendrogram obtained by combined random amplification of polymorphic DNA patterns of the isolates from 10 days of daily sourdough propagation at 30°C for 6 h by using wheat flour type 0 F114. Primers P4, P7, and M13 were used for RAPD-PCR analysis. Lactic acid bacterium isolates I(1) to I(100) correspond to those of Table 1. Cluster analysis was based on the simple matching coefficient and unweighted-pair group method with arithmetic average.

To confirm the data from RAPD-PCR, REA-PFGE analysis was carried out for the starters L. sanfranciscensis LS8, LS14, and LS44 and the isolates found in sourdoughs S-LS8, S-LS14, and S-LS44. The PFGE fingerprints of SmaI and ApaI digests showed that profiles of the isolates I 31-32, I 52 and I 54-55, and I 81-83 were identical to those of starters LS8, LS14, and LS44, respectively (data not shown).

Acidification during sourdough propagation.

During sourdough propagation, the median values of ΔpH ranged from 1.81 (control) to 2.13 (S-LS12) (Fig. 4). The acidifying activity (decrease of pH and organic acids production) of the control increased during propagation, and it was almost constant after the fourth day (ΔpH, 1.84). The acidifying activity between the nine started sourdoughs was rather homogeneous. Only S-LS6 and S-LS48 showed median values slightly lower than those of the other sourdoughs. All started sourdoughs showed differences between the 5th and 95th percentiles of the aggregate data of ≤0.2 units. The concentration of d,l-lactic acid varied from about 32 (control at the first day) to 70 mM (S-LS44 after 10 days). The concentrations of d,l-lactic acid of sourdoughs S-LS3, S-LS8, S-LS12, S-LS14, and S-LS44 were significantly (P < 0.05) constant throughout propagation. By comparing the first day to the 10th day of propagation, the concentrations of d,l-lactic acid for S-LS6, S-LS40, S-L41, and S-LS48 significantly (P < 0.05) increased by 10 mM or more. During propagation, the concentration of acetic acid varied from about 5 (control at the first day) to 18 mM (S-LS6 after 10 days). After 10 days, only S-LS41 (ca. 16 to 8 mM) and S-LS48 (ca. 14 to 6 mM) showed significant (P < 0.05) decreases compared to results for the first day of propagation. The variations in the concentration of d,l-lactic and/or acetic acids determined variations in the quotient of fermentation (QF). Only the QF of sourdoughs S-LS3, S-LS8, S-LS12, S-LS14, and S-LS44 was significantly (P < 0.05) constant throughout propagation.

FIG. 4.

ΔpH values (difference in pH units between the initial pH and final pH after sourdough propagation at 30°C for 6 h during 10 subsequent days by using wheat flour type 0 F114. Data are the means from three independent experiments (n = 3). The center line of each box represents the median (□); the top and bottom of the box represent the 75th and 25th percentiles of the data, respectively. The top and bottom of the error bars represent the 5th and 95th percentiles of the data, respectively. The circles in each box plot extend to the outliers of the data (○). Control, unstarted sourdough. Sourdough started with Lactobacillus sanfranciscensis LS3, LS6, LS8, LS12, LS14, LS40, LS41, LS44, or LS48 is indicated.

Proteolysis during sourdough propagation.

The total concentration of FAA was determined throughout propagation. The control showed the largest variation when results for the first (ca. 68.9 mg kg⁻¹) and the 10th (140.3 mg kg⁻¹) days of propagation were compared. S-LS8, S-LS14, and S-LS44 maintained an almost constant concentration of FAA during propagation (170.4 to 171.7, 110.5 to 120.9, and 152.1 to 156.8 mg kg⁻¹, respectively). Asp, Leu, and Orn were the FAA which showed the largest variations. The other sourdoughs had variations of FAA intermediate between those for the control and the above samples (data not shown).

CLCPs.

The catabolic profiles of the microbiota of sourdoughs during propagation were determined by calculating the indices H′, S, and E (Table 2). According to the utilization pattern substrate (H′ index), sourdoughs were significantly (P < 0.05) distinguished in several groups: (i) control, showing a decrease after the fifth day of propagation; (ii) S-LS3, S-LS6, and S-LS40, characterized by the opposite variation during time; (iii) S-LS41 and S-LS48, showing a decrease until the fifth day of propagation was reached; and (iv) S-LS8, S-LS12, S-LS14, and S-LS44, showing a decrease throughout propagation. Although all sourdoughs had similar values of the H′ index at 1 day, sourdoughs harboring persistent L. sanfranciscensis LS8, LS14, and LS44 had the lowest H′ indices after 10 days of propagation. The S index had the same trend of H′ and grouped sourdoughs similarly. The E index gave a measure of the statistical significance (equitability) of the values expressed by the H′ and S indices.

TABLE 2.

CLCPs of sourdough microbiota^a

Sourdough and timeb	Value for indexc
	H′	S	E
Control
T1	2.98 ± 0.12	14.70 ± 2.08	2.55 ± 0.03
T5	3.02 ± 0.12	14.33 ± 1.53	2.60 ± 0.03
T10	1.27 ± 0.17	3.33 ± 1.52	2.44 ± 0.83
S-LS3
T1	3.24 ± 0.06	19.33 ± 1.15	2.51 ± 0.09
T5	1.33 ± 0.44	2.50 ± 0.70	3.33 ± 0.36
T10	3.01 ± 0.11	18.93 ± 2.51	2.35 ± 0.05
S-LS6
T1	3.27 ± 0.02	19.7 ± 1.53	2.53 ± 0.08
T5	1.83 ± 0.13	6.00 ± 1.00	2.35 ± 0.13
T10	2.90 ± 0.09	16.67 ± 1.53	2.38 ± 0.04
S-LS8
T1	2.52 ± 0.26	11.67 ± 2.08	2.35 ± 0.12
T5	1.57 ± 0.01	5.00 ± 0.05	2.24 ± 1.19
T10	1.02 ± 0.10	3.00 ± 1.00	2.13 ± 0.026
S-LS12
T1	3.13 ± 0.04	17.00 ± 0	2.54 ± 0.03
T5	2.65 ± 0.07	12.66 ± 1.52	2.41 ± 0.12
T10	2.08 ± 0.49	8.33 ± 5.03	2.26 ± 0.11
S-LS14
T1	3.07 ± 0.13	16.33 ± 0.57	2.54 ± 0.14
T5	2.66 ± 0.18	12.66 ± 0.57	2.42 ± 0.13
T10	1.04 ± 0.04	4.33 ± 0.57	1.63 ± 0.07
S-LS40
T1	3.29 ± 0.03	20.67 ± 1.15	2.49 ± 0.4
T5	1.63 ± 0.21	5.00 ± 1.00	2.33 ± 0.14
T10	2.99 ± 0.13	16.00 ± 0	2.49 ± 0.11
S-LS41
T1	3.21 ± 0.06	19.67 ± 1.53	2.49 ± 0.11
T5	2.60 ± 0.07	12.70 ± 1.53	2.36 ± 0.13
T10	2.58 ± 0.16	12.50 ± 1.00	2.35 ± 0.12
S-LS44
T1	3.24 ± 0.06	19.00 ± 1.73	2.53 ± 0.11
T5	1.88 ± 0.13	11.50 ± 0.12	1.77 ± 0.39
T10	0.93 ± 0.23	2.67 ± 0.57	2.16 ± 0.62
S-LS48
T1	3.06 ± 0.01	18.64 ± 2.96	2.41 ± 0.88
T5	2.99 ± 0.04	15.33 ± 1.15	2.51 ± 0.10
T10	2.99 ± 0.48	16.00 ± 2.83	2.16 ± 0.06

^aSourdough microbiota was profiled after one (T₁), five (T₅), or 10 (T₁₀) days of daily sourdough propagation at 30°C for 6 h with wheat flour type 0 F114. Values are means ± standard deviations (n = 3).

^bControl, unstarted sourdough. Sourdough (S) started with Lactobacillus sanfranciscensis LS3, LS6, LS8, LS12, LS14, LS40, LS41, LS44, or LS48 is indicated.

^cH′, Shannon's diversity index; S, substrate richness; E, substrate evenness.

Further use of persistent Lactobacillus sanfranciscensis strains.

In order to investigate the performances of the persistent starters in fermentations using a different flour, the same protocols for sourdough production and propagation were repeated for the three persistent L. sanfranciscensis strains (LS8, LS14, and LS44) by using another wheat flour, type 0 F164. After 8 h of fermentation at 30°C (sourdough production), all started sourdoughs had cell densities of presumptive lactic acid bacteria like those found with wheat flour type 0 F114. On the contrary, the control had values of about 7.7 log CFU g⁻¹. During propagation, the median value of presumptive lactic acid bacteria in the control was about 9.4 log CFU g⁻¹. The median values for started sourdoughs were found to be 9.6, 9.8, and 9.8 log CFU g⁻¹ for S-LS8, S-LS14, and S-LS44, respectively. The ranges of the 5th and 9th percentiles were also very narrow when flour type 0 F164 was used. The dominant microbiota of the raw wheat flour type 0 F164 was less complex than that of wheat flour type 0 F114. The following species were identified: W. confusa (17 isolates), L. sanfranciscensis (6 isolates), L. rossiae (4 isolates), and Lactobacillus spp. (3 isolates). Table 3 and Fig. 5A and B show the species and strains identified after 10 days of propagation. As determined by RADP-PCR and REA-PFGE analyses, the three starters persisted during this propagation also (data not shown). The control, S-LS8, and S-LS44 contained the association between W. confusa and L. sanfranciscensis. S-LS14 also contained L. rossiae. Also, in this case, several strains belonging to the same species were the same in different sourdoughs (e.g., strains of W. confusa in the control and S-LS8 or in the control and S-LS14 and strains of L. sanfranciscensis in the control and S-LS44 or in S-LS14 and S-LS44). Apart from the starter used, five different strains of L. sanfranciscensis variously emerged as dominant in sourdoughs.

TABLE 3.

Identification of sourdough lactic acid bacteria by using 16S rRNA gene sequencing^a

Sourdoughb	Isolatesc	Closet relative(s)	Accession no. of closest relative (identity ≥ 98%)
Control	I(101), I(102)	W. cibaria/ W. confusad	AJ295989/AB023241
	I (103), I(104)	W. cibaria/ W. confusad	AJ295989/AB023241
	I (105)-I(108)	L. sanfranciscensis	EU350220
	I (109), I(110)	L. sanfranciscensis	EU350220
S-LS8	I (111), I(117)-I(119)	W. cibaria/ W. confusad	AJ295989/AB023241
	I (114), I(115)	W. cibaria/ W. confusad	AJ295989/AB023241
	I (112), I(113), I(116)	L. sanfranciscensis	EU350220
	I (120)	L. sanfranciscensis	EU350220
S-LS14	I (121), I(126)-I(128), I(130)	W. cibaria/ W. confusad	AJ295989/AB023241
	I (122), I(123)	L. sanfranciscensis	EU350220
	I (125)	L. sanfranciscensis	EU350220
	I (124)	L. rossiae	AJ564009
	I (129)	L. sanfranciscensis	EU350220
S-LS44	I (131), I(137)	L. sanfranciscensis	EU350220
	I (134)	L. sanfranciscensis	EU350220
	I (139), I(140)	L. sanfranciscensis	EU350220
	I (132), I(133), I(135), I(136)	W. cibaria/W. confusad	AJ295989/AB023241
	I (138)	L. sanfranciscensis	EU350220

^aBacteria were isolated after 10 days of daily sourdough propagation at 30°C for 6 h with wheat flour type 0 F164.

^bLS8, LS14, and LS44 refer to strains of Lactobacillus sanfranciscensis used as the starters.

^cIsolates with different RAPD-PCR profiles are grouped in separate rows.

^dIsolates were allotted to W. confusa by comparing their DGGE profiles obtained after rpoB gene PCR with the corresponding profiles of the type strains.

FIG. 5.

(A) Representative RAPD-PCR patterns of lactic acid bacteria isolated after 10 days of daily sourdough propagation at 30°C for 6 h by using wheat flour type 0 F164. Primer P7 was used for RAPD-PCR analysis. st, DNA molecular size standards (12,000 to 100 bp); Control, unstarted sourdough. Sourdoughs started with Lactobacillus sanfranciscensis LS8 (S-LS8), LS14 (S-LS14), or LS44 (S-LS44) are shown. Lactic acid bacterium isolates (I) are coded based on partial 16S rRNA and recA gene sequence comparisons and correspond to those of Table 3. Lb., Lactobacillus. (B) Dendrogram obtained by combined random amplification of polymorphic DNA patterns for the isolates from 10 days of daily sourdough propagation at 30°C for 6 h by using wheat flour type 0 F164. Primers P4, P7, and M13 were used for RAPD-PCR analysis. Lactic acid bacterium isolates I(101) to I(140) correspond to those of Table 3. Cluster analysis was based on the simple matching coefficient and unweighted-pair group method with arithmetic average.

During propagation, acidification, QF, and proteolysis of the control and started sourdoughs did not significantly (P < 0.05) differ from those reported when wheat flour type 0 F114 was used.

DISCUSSION

L. sanfranciscensis very often emerges as the natural inhabitant of mature sourdough type I (5, 18, 21, 23, 25, 29, 32). Since it possesses interesting technological and functional properties (12, 19), L. sanfranciscensis is considered one of the most promising candidates to be used as a starter for sourdough fermentation. First, this study involved describing the taxonomic structure and monitoring the dominant population of lactic acid bacteria during wheat flour sourdough type I propagation by using nine starter cultures of L. sanfranciscensis.

Since the first day of propagation, all started sourdoughs reached cell numbers of presumptive lactic acid bacteria that are usually found in mature sourdoughs (18). During propagation, cell numbers varied slightly between and within starters, thus indicating a quantitatively stable microbial composition. After the third day of propagation, the control sourdough, without starter added, reached almost the same stable cell numbers. Accordingly, acidification of singly started sourdough was almost constant during propagation, and differences among sourdoughs were very limited.

As shown by other authors (38), the establishment of a stable population of lactic acid bacteria occurred through a three-phase evolution within a week. As determined by RAPD-PCR and REA-PFGE analyses, only three (LS8, LS14, and LS44) of the nine starters used dominated throughout 10 days of propagation carried out under rigorously standardized conditions. The others were outcompeted by the autochthonous population of the wheat flour and disappeared progressively starting from the first day of propagation. Regarding autochthonous lactic acid bacteria, all the other dominant strains, except for L. brevis, found at the end of propagation corresponded to species isolated from raw wheat flour type 0 F114. One autochthonous strain of L. sanfranciscensis was found to be dominant in all sourdoughs. Persistent starters were variously associated with strains of L. sanfranciscensis (starter LS8), L. sanfranciscensis and W. confusa (starter LS14), or L. sanfranciscensis and L. plantarum (starter LS44). Most of the autochthonous species (originating from wheat flour), L. paralimentarius, L. rossiae, L. lactis subsp. lactis, and P. pentosaceus, were found only in sourdoughs where L. sanfranciscensis starters disappeared. Other research studied the microbial stability of sourdoughs, especially type I. L. sanfranciscensis, L. paralimentarius, L. plantarum, and Lactobacillus pontis dominated the lactic acid bacteria population in Belgian type I sourdoughs (32). Species belonging to the genera Lactobacillus, Pediococcus, and Weissella were found to be dominant in sourdoughs produced under laboratory-scale conditions (38). L. sanfranciscensis suddenly became dominant during dual bacterial combinations with sourdough subdominant species (7, 23) and during conventional and organic rye flour fermentations (29). Nevertheless, only one study (25) previously considered four sourdoughs produced under practical conditions by using a starter mixture of three commercially available sourdough starters. Also, in this case, among several lactic acid bacterium species contained in the starter mixture, only a few Lactobacillus strains competed under the prevailing ecological conditions, becoming dominant within a remarkably short time. In the current study, the three persistent starter strains, LS8, LS14, and LS44, were further used for sourdough production and propagation by using another wheat flour, type 0 F164. The endogenous population of lactic acid bacteria of this flour partly differed from that characterizing the previous one, but also in this case all three strains persisted throughout propagation. Under these conditions, stable associations were always found, mainly with other strains of L. sanfranciscensis and W. confusa.

Largely different autochthonous strains of L. sanfranciscensis became dominant when wheat flour type 0 F164 was used. The population composition of lactic acid bacteria in traditional Belgian sourdoughs had been previously reported to be influenced by the bakery environment rather than by the type of flour (31). A further study (32), carried out with traditional Belgian sourdoughs sampled at 11 bakeries, showed little temporal microbial and metabolic variations despite the use of different flour batches and possible variations in flour characteristics during long-time propagation. Nevertheless, the above studies considered the stability of mature sourdoughs, wherein microbial consortia had been selected during a short time, thus developing a high level of stability toward endogenous and exogenous parameters (14). Although this study was carried out under laboratory conditions that might have differed from bakery environmental conditions, the following was shown about wheat flour: (i) it is the source of autochthonous lactic acid bacteria that can associate with or outcompete starter lactic acid bacteria, and (ii) it plays a key role in establishing the stable microbial consortia within a short time, probably due to composition in fermentable substrates, nutrients, growth factors, minerals, buffering capacity, and efficacy of growth-inhibiting principles (21). Apart from the starter used, all sourdoughs propagated by using the two types of wheat flour harbored L. sanfranciscensis strains. Overall, the stable persistence of L. sanfranciscensis in sourdough may be explained through the following: (i) a unique central metabolism and/or transport of sourdough-specific carbohydrates (15), (ii) activated proteolytic enzymes and/or an arginine deiminase pathway (9, 17), (iii) particular stress adaptation responses (10), (iv) well-developed mechanisms of interspecies communication (16), and (v) synthesis of antimicrobial compounds (15).

The metabolic profiles of the sourdough lactic acid bacteria are likely to vary according to changes in the microbial population during propagation and regardless of starter persistence. First, this study determined the CLCPs as an estimation of the global metabolic activities of the sourdough communities. Since it is highly reproducible, this approach is largely used to study the functional diversity of other complex ecosystems, such as soil microorganisms (20, 22, 28, 40). The results of this study implied that the catabolic response profiles of compounds belonging to several chemical classes (carbohydrates, carboxylic acids, polymers, amino acids, amines, and miscellaneous substrates) were linked to the microbial diversity of sourdough. Catabolic profiles of sourdoughs containing persistent starter cultures of L. sanfranciscensis behaved similarly during propagation and were clearly differentiated from those of the other sourdoughs. Almost the same differentiation between sourdoughs was found by determining the QF and the concentration of FAA.

The use of starters in sourdough fermentation could be an attractive perspective. Obviously, starters have to be selected based on technology and functional features, but as shown in this study, only a few of them may persist and compete with autochthonous strains within short-time propagation and before the stable microbial consortium is established. Sourdough cell numbers and acidification may reach optimal values, but they do not guarantee the persistence of starters. Strain robustness toward environmental conditions and microbial competitors has to be considered as one of the main criteria for selecting L. sanfranciscensis starters to be used once a week or probably monthly under daily back-slopping.